High nutritional risk by three variants of NUTRIC score in critically ventilated patients

Abstract

Introduction: variants of the NUTRIC score with or without inflammatory biomarkers, modified without interleukin 6 (IL-6) (NUTRICm), with C-reactive protein (CRP) instead of IL-6, dichotomous (NUTRICpcr1) or in tertiles (NUTRICpcr2), were proposed to assess nutritional risk (NR) in critical patients. However, the assessment of the high NR might not be uniform between these scores. Objectives: to compare the assessment of the high NR by NUTRICm and the two variants of the NUTRICpcr. Material and methods: analysis of a prospective cohort of patients ventilated prior to COVID-19. Agreement was analyzed using the Kappa test and mortality discrimination by logistic regression. The proportion of patients with high NR was compared with the Chi-square test. Results: five hundred and fifty patients were analyzed. Median (IQR) age and APACHE II: 44 (28-58) years and 17 (12-22) points, predominant traumatic pathology (38.2 %) and Intensive Care Unit (ICU) mortality of 32.5 %. The concordance was high between NUTRICm and NUTRICpcr1 (Kappa = 0.81) and lower between NUTRICm and NUTRICpcr2 (Kappa = 0.60). The AUCROC (95 % CI) of NUTRICm, NUTRICpcr1 and NUTRICpcr2 to discriminate mortality was 0.695 (0.495-0.591), 0.693 (0.495-0.591) and 0.685 (0.495-0.591), respectively. The tall NB showed significant differences between NUTRICm and NUTRICpcr1 (19.8 % vs 14.4 %, p 0.0243), being greater between NUTRICm and NUTRICpcr2 (19.8 vs 9.8 %, p < 0.0001). Conclusion: the three NUTRIC variants studied discriminate mortality in a similar way. However, the NUTRICm, without an inflammatory biomarker, classifies more patients as high nutritional risk.